Professional Documents
Culture Documents
CONTENTS
Foreword Chapter 1: Introduction ................................................................................................................................................................. 1
Designing effective policy ............................................................................................................................................... 1 Who is this guidance for? ................................................................................................................................................. 1 Whats the point? ................................................................................................................................................................ 1
Annex A: The Rationale for Government Intervention .......................................................... 17 Annex B: Some Additional Guidance on Valuing Costs and Benefits ............ 18 Annex C: Where Can I Get Help? ................................................................................................................................... 19 Glossary ................................................................................................................................................................................................................. 20
Chapter 1: INTRODUCTION
Designing Effective Policy
Appraisal and evaluation are vital components of sound policy making. They ensure that our actions are well thought through and are framed by evidence on what works and what does not. When we consider developing new proposals, we first need to be very clear why we are considering doing something. What is wrong with the status quo? What is the rationale for government intervention? We then need to be equally clear about what we want to achieve, specifying some very precise and measurable objectives. But what are appraisal and evaluation? Appraisal is the process of examining options for meeting policy objectives and weighing up the costs, benefits, risks and uncertainties of these options before a decision is made. Evaluation is a retrospective analysis of a policy to assess how successful or otherwise it has been and what lessons can be learnt for the future. For both appraisal and evaluation, what counts when judging the success of a policy is the difference it made compared to what would have happened instead. Establishing what would have happened instead is often the hardest bit, and the use of pilots, and particularly the use of control groups1, can be very important. This guidance will help steer you through the key considerations you must make when developing policy. The information is based on HM Treasurys Green Book which constitutes binding guidance for government departments.
Effectiveness having the largest impact possible on the objectives set Efficiency achieving the best outcome or maximum output in relation to the resources used Economy ensuring that the resources used are of best value
Put simply, appraisal and evaluation help us ensure that the Departments policies meet our commitment to get the best results possible from public money.
1. Similar individuals or people who are not offered the policy or programme
Why do we need to do anything about this? What is the justification for us acting or spending public money? What outcome are we hoping to achieve? What are the different ways we can achieve that outcome, and which is the best way? Have we sufficiently considered the risks of things going wrong? What are going to be the impacts and costs on different groups within society? How are we going to know whether we are being successful? What are we going to do to learn from this?
The Departments standard approach is to consider these questions in the ROAMEF framework, which is illustrated in Figure 1. Figure 1: The ROAMEF Policy Making Cycle
Rationale
Objectives Feedback
Appraisal
Evaluation
Considering the ROAMEF framework, effective policy making starts with a clear, evidence-based statement of the Rationale for DfES intervention; what is the problem our intervention aims to solve? Are we sure that DfES intervention is the best way to solve the problem? Out of this clear understanding of the problem should flow specific and measurable Objectives that the DfES intervention aims to achieve. We use Appraisal to consider the various options of how to meet these identified policy objectives. This is a key aid to selecting the best course of action. Policy implementation follows, and during this stage we Monitor the progress and effectiveness of the intervention.
Policies should be subject to an Evaluation, the planning for which should precede policy implementation. The evaluation results should offer Feedback on the success of the policy which can inform subsequent policy decisions, and add to the DfESs knowledge base on what works and what does not.
How central the initiative is to the Departments objectives and broader Government policy aims The overall scale of expenditure on the policy Whether the appraisal or evaluation results will add significantly to our existing knowledge of what works Whether the evaluations results will be used to redesign or change the policy or project
You should also consider whether there is scope for joining up evaluations of policies across your directorate or indeed across the department.
Insufficient investment in certain types of education or skills to secure economic prosperity, international competitiveness or societys well-being? Some would call this economic inefficiency or market failure. For example, is there evidence that the economy or society is worse off due to a lack of particular skills in the workforce? Is there evidence that our attainment or participation levels lag behind those of our international competitors? A loss of fairness, or unequal access to some types of education? For example, is there evidence that because some individuals are born into poor households, they miss out on the good start that nursery education affords?
If the answer to either or any similar question is yes, there may be justification for intervention, provided that the likely benefits of doing something exceed the costs. Rationale for policy should be evidence-based. The departments Research Portal2 is an excellent tool for marshalling pertinent research evidence and you can also seek advice from specialists in your directorate. Annex A offers more general background on the rationale for government intervention.
Objectives
The next step is to define our objectives. Understanding WHY we want to intervene (i.e. establishing the rationale) helps us understand WHAT it is that we are trying to achieve by that intervention. It helps us to clarify exactly what difference to outcomes any policy or project is intended to make, over and above what would have happened if we were not to intervene. In specifying policy objectives, it is instructive to think of a hierarchy of ultimate outcomes (for society), outputs (of the education system or from childrens social services) and inputs (into the system) as shown in Figure 2. There will usually be processes that turn inputs into outputs whilst outputs will ultimately translate into outcomes for society. Figure 2: Outcomes, Outputs and Inputs
Inputs include funding for more early years education places; increased teachers pay; connecting every school to the internet; improving support for foster carers
Processes include better teaching; better availability to information for young people and parents; better incentives to learn or stay on
Outputs include increased pupil knowledge; better early cognitive ability; better adult literacy and numeracy; greater post-16 participation; better workforce skills; more adult learners
Outcomes include reaised economic growth; higher employment; greater fairness or equality; better health or lower crime
2. The research Portal is available through Pathfinder on PRISM. The weblink is http://ntweb1/researchportal/Homepage.asp
HM Treasury will want us to justify our policies in economic terms, which means having a view of the outcomes3 which education and skills policy ultimately contributes to. It is often difficult to specify policy objectives in terms of outcomes, so we often specify them in terms of outputs. Policy objectives should not be specified solely in terms of inputs or processes (although it will often be necessary to monitor these to ensure resources are being used properly or that political commitments are being met). We are ultimately judged on the outputs we produce, and the outcomes these lead to, not the inputs we put into the system or the processes that lead to outputs. In all cases, objectives should be defined in such a way that it can be easily established after the event whether, and to what extent, they were met. Good objectives should be SMART: Specific Measurable Agreed Realistic Time-bound Finally, policy objectives should relate to the aims and objectives in the DfESs Five Year Strategy document published in July 2004, and to Public Service Agreement (PSA) targets. Indeed, your rationale and objectives statement (see below) should make reference to the DfES objectives or PSA targets your policy is furthering. See Annex C for web links to the DfES Five Year Strategy and PSA targets.
the rationale for the policy, citing relevant evidence, and the objectives for the policy, which should be SMART
All policies should have a Rationale and Objectives statement drawn up at the outset by the responsible policy team. Your analytical contacts will be able to advise you as you draft this document. The Research Portal is an excellent aide in marshalling evidence for the policy rationale, and is available through Pathfinder on PRISM. The Rationale and Objectives statement is a helpful tool in the early stages of the policy development process. It is an essential starting point for developing policy options, appraising these options and evaluating a policy.
3. See the DfES publication Education and Skills: The Economic Benefit, pages 13 to 22 for a graphical introduction to the evidence on the contribution of education and skills to ultimate outcomes. The publication is available at www.dfes.gov.uk/economicbenefit
Check
Reference
Pages 4-5
Pages 7-8
Pages 9-10
Page 9
Page 11
Pages 11-12
In some cases it might appear that there is only one policy option and that appraising this option or considering alternatives is not going to add anything useful to the foregone decision. This approach should be avoided. Appraisal can often reveal that there are better alternatives, or serious risks to a policy. At the very least, there is always the do nothing option, which must be considered as seriously as options to act or spend public funds. Even where time is short, it is better to start with a full range of options. In drawing up options, you should consider the evidence on how various policy levers or options might lead us towards the objectives. You should also consult as widely as possible in drawing up the range of options. Each option should be assessed against the do nothing option in terms of: impact on objectives acceptability and practicality likely costs and benefits Out of this outline assessment a shortlist of options may be created in order to keep the appraisal process manageable. However be aware of the risk that a process of short-listing may eliminate the optimal solution before it is given full consideration. Try to ensure that your shortlist still covers a wide range of potential action and make sure you record the reasons for rejecting an option.
Benefits
Costs
Direct Costs (one-off ) e.g. policy implementation costs; cost of any buildings or equipment Direct Costs (ongoing) e.g. tuition costs; policy admin costs Indirect Costs e.g. earnings foregone by students whilst participating in post-16 study
Higher early cognitive ability Raised school attainment Better skilled adults Improved behaviour Improved child outcomes e.g. health or crime Reduced child abuse Higher employment rates More productive workers
It is important to isolate the additional benefits and additional costs. The word additional refers to an impact over and above that which would have happened under the do nothing option. In other words, the additional benefits of an option are its benefits less the benefits of the do nothing option; additional costs is similarly defined. Note that the benefits and costs of the do nothing option are not generally nothing and are not necessarily those that were occurring before the policy started. Using control groups or taking background factors into account are important in isolating additional benefits. Specialists in your directorate will be able to advise on how to best estimate additional benefits and costs in the case of your policy.
Discounting
Most appraisals will have to compare costs and benefits which occur in different time periods (typically measured in years). Normally people prefer to receive benefits such as goods and services sooner rather than later and to pay bills or costs later rather than sooner. In other words they value 1,000 more if they receive it today than if they receive it in 10 years time. So when conducting appraisals, to make all costs and benefits comparable regardless of the year in which they occur, they need to be converted to present values Converting into present values is also called . discounting, and is done by multiplying costs and benefits by a discount factor (see next paragraph). Discounting should not be confused with any adjustment for the impact of inflation. The discount factor is calculated from the discount rate, which is defined as the rate at which the value of a pound held today falls away over time. The Treasury requires that a discount rate of 3.5% per annum is used in appraisals. At a discount rate of 3.5% or 0.035, the discount factor for a given year t, is given by the formula: Discount factor in year t = 1/(1+0.035)t So the discount factor in year 5 with a discount rate of 3.5% is 1/(1+0.035)5 which equals 0.842. The box below gives an example of how to discount a stream of benefits to give their present values.
Costs
Additional Benefits () Discount Factor Discounted Benefits ()
0
1,000 1 1,000
1
1,000 0.966 966
2
1,000 0.934 934
3
1,000 0.902 902
4
1,000 0.871 871
5
1,000 0.842 842
The discounted benefits can be interpreted as follows. Because we prefer receiving benefits sooner rather than later, we would be equally happy receiving a smaller benefit of 842 today as one of 1,000 in 5 years time. So, 1,000 of benefits in 5 years time is valued the same as 842 of benefit today, and 1,000 of benefits in 3 years time is valued the same as 902 of benefits today, and so forth.
4. Note that by convention, the year in which a policy starts is called year 0, and subsequent years are year1, tear 2 etc. 5. Real benefits are benefits adjusted for the impact of inflation 6. For costs and benefits more than 30 years into the future, the Green Book (pages 98-99) says lower discount rates should be used. A link to the green Book is at Annex C.
The options may have varying impacts on people of different incomes, ages, locations, ethnic origins or genders. The Departments intention to mainstream equal opportunities means that the distributional impact of proposed policies needs to be seriously considered as part of your appraisal7. Even if two options have exactly the same net benefits, if one creates a more equitable outcome, it will normally be justifiable to prefer it on that basis. The Treasury asks that a distributional adjustment is factored into the NPV calculation itself if (a) the proposal has significant distributional impacts and (b) it is possible to sensibly estimate the impact on different outcome groups. It is recommended that you seek specialist advice before attempting this.
Net Present Value (NPV) and Internal Rate of Return (IPP) Calculations: A worked example
Assume we have a policy objective to raise GCSE attainment. Two policy options are being considered, both of which cover a five year period. Option 1, a policy to raise teaching standards, has initial costs of 500 in years 0 and 1, and produces additional benefits of 300 in years 1 and 2 and of 500 in years 3 and 4. Option 2, a policy to reduce truancy, has initial costs of 700 in year 0 and 500 in year 1. It produces additional benefits of 400 a year from year 1 to year 4. The costs, benefits, net benefits and discounted net benefits of each option are shown in Table 3 Table 3: NPV Calculation example (All figures in s)
Year
Option 1 Benefits Costs Net Benefits Discounted Net Benefits Option2 Benefits Costs Net Benefits Discounted Net Benefits
NPV
IRR
474
20%
286
11%
The net benefits row is calculated as benefits less costs. The discounted net benefits row is calculated as the net benefits for a given year multiplied by the discount factor for that year8. NPV is the sum of all the discounted net benefits. As shown in Table 3, Option 1 has an NPV of 474 and Option 2 has an NPV of 286. The (internal) rates of return for each option are 20% and 11% respectively. As both NPVs are positive or equivalently both rates of return are above 3.5% - both options meet the Treasurys criterion for investment of public funds. However, since Option 1 has a greater NPV than Option 2, it would be preferred to Option 2, all things equal. The expression all things equal is important. Where options involve different costs to government, NPV estimates may need to be adjusted to reflect this. Analysts in your directorate can advise here. On a practical note, Microsoft Excel is a good programme to use to calculate NPVs and Internal Rates of Return (IRR), since it has functions that will compute these two measures for a row or column of numbers. To access the NPV or IRR functions click on the f* button in Excel and type NPV or IRR in the search box.
7. You will find a lot of useful questions to consider and advice on the Fairway site http://ntweb1/newmainstreamingwebtool/default.htm 8. Discount Factors are shown in Table 2
10
11
12
Check
Reference
Pages 4-5
Pages 13-14
Page 14
Page 15
Pages 15-16
What is to be Evaluated?
You need to clearly specify the questions the evaluation aims to answer. Ideally these should be focused on identifying the outcomes of the policy, and explaining how they came about. Evaluation focused on the impact of the project on outcomes is called summative evaluation. Summative evaluation should consider additional impact at all times (see additional benefits on page 8). For example, summative evaluation would consider whether a policy such as Specialist Schools had an additional impact on pupils attainment and attendance. It would quantify the magnitude of any such impacts. Formative evaluation can be used to determine why, how or under what conditions a policy can be best directed or implemented, in order to shape its future direction. Formative evaluation may consider, for example, whether and why the collaboration element of Specialist Schools was pivotal in instances where
13
Specialist Schools were successful in raising attainment, or in what type of schools collaboration had its greatest effect. Summative and formative evaluation should be complementary. However, as a minimum, all evaluations should include a summative assessment of the overall additional impact of the policy on outcomes. Good practice in performance management means we should continually monitor the progress of a policy towards meeting its objectives. This should involve the systematic collection and analysis of data relating to both the inputs into the policy and early indicators of its outputs. This data is sometimes called monitoring information. Whilst monitoring information can be distinct from data collected for evaluation, policy makers should try to minimise duplication in data collection for monitoring and evaluation purposes.
How it is to be Evaluated
Just as with appraisal, you should identify and measure the direct and indirect benefits and costs of the intervention over and above what would have happened in the absence of the policy. In other words, evaluation should seek to establish the additional impacts of the policy. Where possible the evaluation should include a control group to whom the activity was not applied as this will help clarify the net or additional impact of the project. You should consider how outputs and activities might be measured as part of the evaluation. Some of the data needed will be available through any monitoring information that is being collected, or through existing DfES data sources such as the National Pupil Database. Some additional data or evidence may need to be specifically collected for the evaluation.
The questions which the evaluation seeks to answer The evaluations likely data requirements Potential evaluation methodologies Outputs expected from the evaluation The timescale within which the evaluation should report Evaluation costs
The first bullet above will typically include questions concerning the extent to which the policy has met the policy objectives as set out in the Rationale and Objectives statement (see pages 4-5). The EP is a good basis for any subsequent Invitations to Tender. You should seek advice from specialists in your directorate in drawing up the EP.
14
drafting Evaluation Plans, and any Invitations to Tender deciding the evaluations methodology and data requirements tendering and drawing up evaluation contracts commenting on interim reports variations to contract over the course of the research publication and presentation of results
How effective the activity was in achieving its objectives, and why Whether the activity has been more beneficial for some groups than others, and why Suggestions for how the policy could be improved to make it more effective The cost effectiveness of the activity if practicable, a brief discussion of cost-effectiveness relative to other ways of achieving the objectives is useful Why the outturn may have differed from that foreseen in the appraisal What the results imply for future management or policy decisions
The results should generally lead to recommendations for the future. These may include changes in the way the policy is delivered, its cessation or its replacement. An action plan should be developed and executed so that the evaluations recommendations are put into practice. Under the Freedom of Information Act 2000, the Department should publish findings from all the research and evaluation it funds (including policy-funded projects). Many of the departments evaluations are published at www.dfes.gov.uk/research and currently all DfES research and evaluation is published on the last Thursday of every month. For a link to further guidance on DfES publication arrangements, see Annex C. You should make efforts to disseminate evaluation findings widely inside and outside the Department as appropriate. To help you gauge whether you are obtaining good value for money from your evaluation, you should assess, perhaps 6-12 months after publication, the impact your evaluations results have had on policy development in your area and beyond.
9. This is a well-known tool for prioritising risks according to (i) each risks likelihood and (ii) the scale of the adverse impact should each risk happen.
15
To end this chapter, the box below provides a case study of a well-executed evaluation and pilot.
Evaluation Results
The EMA evaluation results estimated that the overall impact of EMA across England would be to raise year 12 participation in full time education by 5.9 percentage points for eligible young people (and 3.8 pp for whole cohort), and year 13 participation by 6.9 percentage points for eligible young people (4.1pp for whole cohort). The design of the evaluation as far as data collected has meant that these results can be disaggregated to see whether EMAs impact has been greater for high versus low attainers, or for high versus low social class individuals. Other breakdowns are possible. The results were made publicly available some evaluation reports are on the DfES research website at www.dfes.gov.uk/research. They can be accessed by typing EMA into the free text search box on this site.
16
17
18
19
GLOSSARY
Additionality Appraisal
An impact arising from an intervention is additional if it would not have occurred in the absence of the intervention The process of defining objectives, examining policy options and weighing up the costs and benefits, risks and uncertainties of those options before a decision is made Data, typically on outcomes, collected immediately before the commencement of a policy, to enable comparison with data collected after the policy has had time to have an effect. A group of individuals who do not receive a policy, but have similar characteristics to those individuals who do receive the policy
Baseline Data
Control Group
Cost Benefit Analysis Comparison of costs and benefits to see if the benefits justify the costs. Cost
Benefit Analysis will typically result in a net present value or a rate of return
Deadweight Discounting Discount Rate Displacement Evaluation Leakage Net Present Value Optimism Bias Regulatory Impact Assessment Sensitivity Analysis Substitution
Expenditure to promote a desired outcome which would have occurred anyway, without the expenditure A method used to convert future costs or benefits to present values using a discount rate The annual percentage rate at which the present value of a future pound is assumed to fall away over time The degree to which increases in the desired outcome are offset by consequential reductions elsewhere Retrospective analysis of a project or policy to assess how successful or otherwise it has been, and what lessons can be learnt for the future Benefits to a policy which accrue to people outside the spatial area or group which was the intended beneficaiary. The discounted value of a stream of future costs and benefits The systematic tendency for appraisers to be optimistic about key project parameters, including capital and operating costs, and benefits delivery An assessment of the impact that regulatory proposals are likely to have on business, charities and the voluntary sector. RIAs consider all available options including non-regulatory alternatives. Analysis of the effects on an appraisals results of varying the projected values of important variables Where a policy creates the desired impact on person A, but at the expense of removing that desired impact from person B
20